Cancer Therapy: Preclinical

Lymphotoxin-b Interacts with Methylated EGFRto Mediate Acquired Resistance to Cetuximab inHead and Neck CancerDennis Shin-Shian Hsu1,Wei-Lun Hwang2, Chiou-Hwa Yuh3, Chen-Hsi Chu4,Yang-Hui Ho5, Pon-Bo Chen6, Han-Syuan Lin3, Hua-Kuo Lin3, Shih-Pei Wu6,Chih-Yi Lin6,Wen-Hao Hsu6, Hsin-Yi Lan6, Hsiao-Jung Wang7, Shyh-Kuan Tai8,Mien-Chie Hung9,10,11,12, and Muh-Hwa Yang1,5,6,13

Abstract

Purpose: In head and neck squamous cell carcinoma(HNSCC), the incidence of RAS mutation, which is the majorcause of cetuximab resistance, is relatively rare compared with theother types of cancers, and the mechanism mediating acquiredresistance is unclear compared with the driver gene mutation–mediated de novo resistance. Here, we investigated the driver genemutation–independent mechanism for cetuximab resistance inHNSCC.

Experimental Design:We used the in vitro-selected and in vivo-selected cetuximab-resistant sublines of HNSCC cell lines forinvestigating the mechanism of acquired resistance to cetuximab.Zebrafish model was applied for evaluating the synergistic effectof combinatory drugs for overcoming cetuximab resistance.

Results: The cetuximab-resistant HNSCC cells undergo a Snail-induced epithelial–mesenchymal transition. Mechanistically,

Snail induces the expression of lymphotoxin-b (LTb), a TNFsuperfamily protein that activates NF-kB, and protein argininemethyltransferase 1 (PRMT1), an arginine methyltransferase thatmethylates EGFR. LTb interactswithmethylated EGFR topromoteits ligand-binding ability and dimerization. Furthermore, LTbactivates theNF-kBpathway through a LTb receptor–independentmechanism. Combination of an EGFR tyrosine kinase inhibitorand a NF-kB inhibitor effectively suppressed cetuximab-resistantHNSCC and interfering with the EGFR–LTb interaction reversesresistance.

Conclusions:Our findings elucidate the mechanism of drivergene mutations–independent mechanism of acquired resis-tance to cetuximab in HNSCC and also provide potentialstrategies for combating cetuximab resistance. Clin Cancer Res;23(15); 4388–401. �2017 AACR.

IntroductionHead and neck squamous cell carcinoma (HNSCC), which

comprises the cancers originating from the oral cavity, orophar-ynx, larynx, and hypopharynx, is one of the major cancers world-wide (1). In recent years, targeting EGFR has become a majorstrategy for combating late-stage HNSCC because EGFR over-expression is observed in approximately 90% of HNSCC casesand is associated with an unfavorable outcome (2, 3). Pivotalclinical studies demonstrated that cetuximab, a humanized IgG1monoclonal antibody against the extracellular domainof EGFR, iseffective in patients with advanced HNSCC when combined withradiation or chemotherapy (4, 5). On the basis of these findings,cetuximab is regarded as a major treatment for advanced orrecurrent/metastatic HNSCC. However, the improvement in sur-vival afforded by cetuximab is not as extensive as expected (6).Notably, although the initial response rate to cetuximab inHNSCC is high,most of the patients eventually develop resistanceto cetuximab, that is, acquired resistance (7). Themechanisms forcetuximab resistance that have been gradually revealed includemutations of RAS or other oncogenic drivers (8, 9), amplificationand activation of ERBB2 (10), deregulation of EGFR cycling andstability (11), and activation of MET (12). We recently demon-strated that methylation of the extracellular domain of EGFRpromotes cetuximab resistance (13). However, although the

1Genome Research Center, National Yang-Ming University, Taipei, Taiwan. 2ThePh.D. Program for Translational Medicine, College of Medical Science andTechnology, Taipei Medical University, Taipei, Taiwan. 3Institute of Molecularand Genome Medicine, National Health Research Institutes, Zhunan, Taiwan.4Department of Life Science and Institute of Bioinformatics and StructuralBiology, College of Life Science, National Tsing Hua University, Hsinchu, Taiwan.5Genomics Research Center, Academia Sinica, Taipei, Taiwan. 6Institute ofClinical Medicine, National Yang-Ming University, Taipei, Taiwan. 7Division ofExperimental Surgery, Department of Surgery, Taipei VeteransGeneral Hospital,Taipei, Taiwan. 8Department of Otolaryngology, Taipei Veterans General Hos-pital, Taipei, Taiwan. 9Department of Molecular and Cellular Oncology, TheUniversity of Texas MD Anderson Cancer Center, Houston, Texas. 10The Uni-versity of Texas Graduate School of Biomedical Sciences at Houston, Houston,Texas. 11Graduate Institute of Cancer Biology and Center for Molecular Medicine,China Medical University, Taichung, Taiwan. 12Department of Biotechnology,Asia University, Taichung, Taiwan. 13Division of Medical Oncology, Departmentof Oncology, Taipei Veterans General Hospital, Taipei, Taiwan.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

D.S.-S. Hsu and W.-L. Hwang contributed equally to this study.

Corresponding Author: Muh-Hwa Yang, Institute of Clinical Medicine, NationalYang-Ming University, No. 155, Sec. 2, Li-Nong Street, Taipei 11221, Taiwan.Phone: 886-228267000 ext. 7911; Fax: 886-228235870; E-mail:mhyang2@vghtpe.gov.tw

doi: 10.1158/1078-0432.CCR-16-1955

�2017 American Association for Cancer Research.

ClinicalCancerResearch

Clin Cancer Res; 23(15) August 1, 20174388

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

KRAS mutation is important for cetuximab resistance, a certainproportion of patients who develop acquired resistance to cetux-imab do not have this mutation (14). Furthermore, the incidenceof RASmutation is relatively low in HNSCC (15), and no clinicalevidence supports the effectiveness of anti-HER2 therapy incetuximab-resistant patients. The inductionmechanism for EGFRmethylation is also unclear. Importantly, the response rate tocetuximab as the second-line treatment in chemotherapy-refrac-tory cases is significantly lower thanwhen cetuximab is used as thefirst-line treatment (16), which implies that a driver mutation–independent cross-resistance mechanism may be involved inresistance to cetuximab.

The epithelial–mesenchymal transition (EMT) is a processwhich enables epithelial cells to acquire mesenchymal properties.EMT is known to participate in several important processesincluding development, fibrosis, and cancer metastasis (17). Abreakthrough in EMT research has revealed that induction of EMTgenerates cells with stem-like properties (18). Because cancer stemcells are responsible for tumor initiation and therapeutic resis-tance, this finding provides crucial links between the acquisitionof metastatic traits, tumor-initiating capability, and treatmentresistance in late-stage cancers. Nevertheless, the role of EMT incancer metastasis is a longstanding debate owing to the difficultyin monitoring this dynamic process in vivo. Recently, two studiesdemonstrated the major role of EMT in chemoresistance ratherthan metastasis (19, 20). For anti-EGFR therapy, EMT has beenshown to regulate thedownstream signal of EGFR (21) andEMT isinvolved in anti-EGFR resistance (22). However, whether EMT-induced anti-EGFR resistance is independent from driver genemutation/amplification–induced resistance and the underlyingmechanismof EMT-induced anti-EGFR resistance remain unclear.

In this study, we demonstrated that the HNSCC cells withacquired resistance to cetuximab undergo a Snail-induced EMT,and lymphotoxin-b (LTb) is crucial for Snail-induced cetuximabresistance. LTb interacts with methylated EGFR to activate theEGFR and NF-kB pathways for mediating cetuximab resistance.Our finding delineates a driver mutation–independent mecha-nism for acquired resistance to cetuximab in HNSCC.

Materials and MethodsCell lines

The human embryonic kidney cell line HEK-293T andthe human hypopharyngeal cancer cell line FaDu were pur-

chased from the Bioresource Collection and Research Centerof Taiwan in 2014. The human oral cancer cell line OECM1was obtained from Dr. Kuo-Wei Chang (National Yang-MingUniversity, Taipei, Taiwan) in 2011. This cell line was orig-inally established by Dr. Ching-Liang Meng of the NationalDefense Medical College in Taiwan (23). The human oralcancer cell lines SAS and CAL-27 were obtained fromDr. Cheng-Chi Chang (National Taiwan University, Taipei,Taiwan) in 2011. The cell lines authentication was performedby examining the DNA-STR profile before establishing theresistant clone and performing relative experiments.

Xenotransplantation into zebrafish embryosThe cells were incubated in CellTrace CFSE (1Degree Bio Inc.)/

PBS working solution (25 mmol/L) for 15 minutes at 37�C andthenwere centrifuged andwashed. The cellswere diluted in PBS to2.3 mL containing 100 tumor cells. A glass transplantation needlewas used to transfer the carboxyfluorescein diacetate succinimidylester (CFSE)-labeled cells into the perivitelline cavity of 2-day-oldzebrafish embryos. After transplantation, embryos were checkedfor the presence of labeled cells 3 hours later. The embryos werethenplaced in a 96-well plate,with 200mLperwell containingonezebrafish embryo and were placed in the incubator. One daypostinjection, the drugs were added into each well. Twentyembryos in 20 wells for each tested drug and the control wereused. The tumor cells were observed under a microscope forproliferation in the zebrafish embryo, and images were takenindividually at 1 day postinjection (1 dpi) and 3 days post-injection (3 dpi). Following image acquisition, embryos werefixed overnight with paraformaldehyde (4%) at room tempera-ture, and immunostaining with a primary antibody against Ki67(1:200, Arigo Biolaboratories Corp.) was then performed to verifyproliferation.

Patient samplesThree sets of HNSCC samples from the Department of

Otolaryngology of Taipei Veterans General Hospital were usedin this study. The first set of samples was applied for analyzingthe expression of LTB in Fig. 2C, which contains tumor sampleswith paired normal counterparts from 12 patients with HNSCCreceiving cetuximab treatment. Among them, four patientsdeveloped recurrent tumors after treatment, and eight patientsdid not have recurrent tumors. The second set of samples,which was obtained from 58 patients with HNSCC beforetreatment, was used to analyze the association between theexpression of LTB and that of SNAI1 (Fig. 2E). The third set ofsamples, which contained samples from two patients whodeveloped a recurrent tumor after cetuximab treatment, wasused for direct sequencing of the driver gene mutations (Sup-plementary Fig. S1G). For these three sets of samples, thecomponents of the tumor tissues were confirmed by pathologicreview before experiments, and most of the tumor specimenswere composed of cancer cells. However, the different ratios ofparenchyma to stroma of each tumor samples may lead to thevariations of the results. In addition to the three sets of samples,we harvested a tumor developed from a cetuximab treatmentpatient for primary culture.

Statistical analysisAn independent-samples t test was performed to compare the

continuous variation of the two groups. A x2 test or Fisher exact

Translational Relevance

The humanized anti-EGFR antibody cetuximab is a majortreatment for advanced head and neck squamous cell carci-noma (HNSCC). In HNSCC, the incidence of RAS mutation,which is the major cause of cetuximab resistance, is relativelyrare compared with the other types of cancers. However, theimprovement in survival after cetuximab treatment is not asextensive as expected, and most patients eventually developacquired resistance despite their initial responses. This reportelucidates the mechanism of driver gene mutations–indepen-dent mechanism of acquired resistance to cetuximab inHNSCC. We also provide potential strategies for combatingthe resistance by blocking the downstream signals in resistantcells or intercepting the EGFR–lymphotoxin-b interaction.

Lymphotoxin-b Induces Cetuximab Resistance

www.aacrjournals.org Clin Cancer Res; 23(15) August 1, 2017 4389

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

OECM1-WT OECM1-CtxR BA

Rank list (x1,000)

159 downregulated genes in EMT

ES: –0.54Nominal P value:0.00FDR q-value:0.11

Res

Sen

0 2010

(ES)

E

60*

FFaDu-Vec

FaDu-Snail

Snail

β-Ac�n

0.40.30.20.10.0

–0.1

ES: 0.41Nominal P value:0.00FDR q-value:0.0

RNA-seq for FaDu-Snail/FaDu-VecGenes ≥ 2.5x, FPKM >1 (n = 477)

Res Sen

Rank list (x1,000)

G

E-ca

dher

in/

N-c

adhe

rin

H

D

87 upregulated genes in EMT 0.5 0.40.30.20.10.0

–0.1

Res

(ES)

Sen

Rank list (x1,000)0 2010

CRe

la�v

e in

hibi

�on

(%)

β-Ac�n

E-cadherin

OECM1

Vimen�n

WT CtxR

0.10.0

–0.1–0.2–0.3–0.4–0.5

Phas

e

40

20

0FaDu-Vec

FaDu-Snail

Cetuximabresistant

Cetuximabsensi�ve

Snail IHC

Rela

�ve

inhi

bi�o

n (%

)

35302520151050

OECM1-WT OECM1-CtxR

sh-scr

sh-Snail

sh-Twist1

**

**

ES: 0.53Nominal P value:0.00FDR q-value:0.09

CtxWt R

Snail

Slug

Twist1

Zeb2

β-Ac�n

Zeb1

OECM1

β-Ac�n

ES: 0.295Nominal P value:0.1845FDR q-value:0.2877

(ES)

Res

Sen

Rank list (x1,000)0 2010

0.250.20.150.10.050.0

–0.05–0.1–0.15

Top 5% upregula�on in FaDu-Twist1 vs. FaDu-Vec (n = 280)

(ES)

FaDu

WT CtxR

Snail

Slug

Twist1

Zeb1

Zeb2

β-Ac�n

Hsu et al.

Clin Cancer Res; 23(15) August 1, 2017 Clinical Cancer Research4390

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

test was applied for comparison of dichotomous variables.The Pearson correlation test was used for analyzing the corre-lation between two continuous factors. For animal studies, nostatistical method was used to predetermine the sample size.The experiments were not randomized. The investigators werenot blinded to allocation during experiments or during out-come assessment.

Accession numbersThe results of the RNA sequencing for FaDu-Snail/FaDu-vec

and the cDNA microarray for OECM1/OECM1-CtxR were depos-ited at Gene ExpressionOmnibus (GEO)with accession numbersGSE65131 and GSE63916, respectively.

Study approvalThis study was approved by the Institutional Review Board of

Taipei Veterans General Hospital (VGHIRB-2016-01-002AC).The animal experiments were approved by the InstitutionalAnimal Care and Utilization Committee of National Yang-MingUniversity (IACUC1041128).

ResultsThe cetuximab-resistant head and neck cancer cells undergo aSnail-induced EMT

To investigate the mechanism of acquired resistance tocetuximab, we generated two cetuximab-resistant HNSCC sub-lines for subsequent experiments. The first cetuximab-resistantsubline (named OECM1-CtxR) was generated by continuoustreatment of the HNSCC cell line OECM1 with cetuximabfor 30 passages for more than 2 months until the cells couldproliferate freely (Supplementary Fig. S1A). Compared withthe OECM1 parental cells, the OECM1-CtxR subline displayeda lower inhibition rate by cetuximab (SupplementaryFig. S1B), and cetuximab had a lesser effect to suppress theEGFR downstream signal pathways ERK and Akt in OECM1-CtxR (Supplementary Fig. S1C). The second cetuximab-resis-tant HNSCC subline was generated by in vivo selection. Thecetuximab-resistant subline was established by harvesting thexenografts for culture from the cetuximab-treated group(named FaDu-CtxR). The control subline was established bycultivating the xenografts from the PBS-treated group (namedFaDu-Ctrl; Supplementary Fig. S1D). The cetuximab resistanceof FaDu-CtxR was confirmed in vivo (Supplementary Fig. S1Eand S1F).

Next, we investigated the characteristics of the generated cetux-imab-resistant HNSCC sublines. We noticed that OECM1-CtxR

cells harbored a mesenchymal phenotype (Fig. 1A and B).The Gene Set Enrichment Analysis (GSEA) showed a signif-icant correlation between the EMT signature and cetuximabresistance (Fig. 1C). We next explored the major EMT regu-lator in this process. We found that among the EMT regula-tors, only Snail was consistently upregulated in both resistantsublines OECM1-CtxR and FaDu-CtxR (Fig. 1D). Knockingdown Snail in OECM1-CtxR attenuated the resistance, whereassuppression of Twist1 did not exert a similar effect (Fig. 1E).Ectopic Snail reduced the growth-inhibitory effect of cetux-imab in FaDu cells (Fig. 1F). GSEA showed that the Snail-upregulated genes in FaDu cells significantly correlated withthe cetuximab resistance signature, whereas overexpression ofTwist1 in FaDu cells did not correlate with the cetuximabresistance signature (Fig. 1G). An increased level of Snail wasnoted in the HNSCC samples with acquired resistanceto cetuximab compared with the cetuximab-sensitive cases(Fig. 1H). These pieces of evidence indicate that Snail-inducedEMT plays an important role in acquired resistance to cetux-imab in HNSCC.

To investigate whether other known mechanisms areinvolved in acquired resistance to cetuximab, direct sequencingof the genes known to be involved in cetuximab resistancewas performed in two resistant sublines (FaDu-CtxR andOECM1-CtxR) and in samples from the recurrent tumors oftwo patients who developed cetuximab resistance after treat-ment. We looked for mutations of PIK3CA (24), EGFR (8),and RAS family genes (9) and for ERBB2 amplification (10, 25).No detectable mutation or ERBB2 amplification was noted(Supplementary Fig. S1G). The expression level of surfaceEGFR, stability of EGFR, cetuximab-binding ability, and phos-phorylated MET levels were similar between the parental cellsand the resistant subline (Supplementary Fig. S1H–S1J). Col-lectively, these data suggest that Snail is crucial in the devel-opment of acquired resistance to cetuximab, which occursindependently of the known driver gene mutations, EGFRexpression, and MET activation.

LTb is critical to mediate Snail-induced acquired resistance tocetuximab and LTb induces EMT

We next explored the major player that mediates Snail-induced cetuximab resistance. First, we obtained the "core sig-nature for cetuximab resistance" by looking at the intersection

Figure 1.The cetuximab-resistant cells undergo Snail-dominant EMT. A, Representative phase-contrast (top) and immunofluorescent images for wild-type OECM1 (OECM-WT) and cetuximab-resistant OECM1 (OECM1-CtxR) cells. Green, E-cadherin; red, N-cadherin; blue, nuclei. Scale bar ¼ 200 mm. B, Western blot analysisof E-cadherin and vimentin in OECM1-WT and OECM1-CtxR cells. C,GSEA result showing the association between the EMT signature and cetuximab resistance (dataobtained from GSE21483). Top, Upregulated gene signature in EMT. Bottom, Downregulated gene signature in EMT. Res, resistance; Sen, sensitive; ES,enrichment scores; FDR, false discovery rate. D,Western blot analysis of the EMT regulators (Snail, Slug, Twist1, Zeb1, Zeb2) in OECM1 versus OECM1-CtxR cells (left)and FaDu versus FaDu-CtxR cells (right). E, Relative growth inhibition by cetuximab of OECM1-WT/OECM1-CtxR cells transfected with the indicated plasmids. Scr, ascrambled control for shRNA experiment. Data represent mean � SD. n ¼ 3. �� , P < 0.01 (Student t test). F, Top, Western blot analysis of Snail in FaDucells stably transfected with a Snail-expressing vector (FaDu-Snail) or a control vector (FaDu-vec). Bottom, relative growth inhibition by cetuximab. The referencesample of each group is the cell line treated with vehicle control. Data represent mean � SD. n ¼ 3. � , P < 0.05 (Student t test). G, Top, GSEA result showing theassociation between the Snail-regulated signature (RNA-seq data from FaDu-Snail/FaDu-vec) and cetuximab resistance. Res, resistance; Sen, sensitive; ES,enrichment scores; FDR, false discovery rate. Bottom, GSEA result for showing the association between the Twist1-regulated signature (FaDu-Twist1 vs.FaDu-CMV; GSE19406) and cetuximab resistance. The top 5% upregulated genes in FaDu-Twist1 (280 identities) were applied as the Twist1-overexpressingsignature. Res, resistance; Sen, sensitive; ES, enrichment scores; FDR, false discovery rate. H, Representative result of Snail immunohistochemistry (IHC) in patientswith cetuximab-sensitive (top; n ¼ 10 cases) versus cetuximab-resistant (bottom; n ¼ 4 cases). Scale bar ¼ 200 mm.

Lymphotoxin-b Induces Cetuximab Resistance

www.aacrjournals.org Clin Cancer Res; 23(15) August 1, 2017 4391

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

121086422468

101214

50403020100

05

10152025

BA

LTβ

β-Ac�n β-Ac�n

WT CtxR

OECM1

LTβ

Ctrl CtxR

FaDu

Cetuximab-treated HNSCC pa�ents

C

r = 0.934P = 0.000

LTB

SNAI1

VGH HNSCC pa�ent (n = 58)E

FaDu-Vec

FaDu-Snail

Snail

LTβ

OECM1-sh-scr

OECM1-sh-Snail

Snail

LTβ

F

r = 0.140P = 0.001

TCGA (n = 542)

G

LTB

SNAI1

D

FaDu-CtxR

50**

sh-scr sh-LTB

Rela

�ve

inhi

bi�o

n (%

)

40302010

0

LTβ

β-Ac�n

sh-scr sh-LTB

50

OECM1-CtxR

**

sh-scr

Rela

�ve

inhi

bi�o

n (%

)

40302010

0sh-LTB

LTβ

β-Ac�n

sh-scr sh-LTB

80Primary HNSCC

sh-scr sh-LTB

Rela

�ve

inhi

bi�o

n (%

)

604020

0

**sh-scr sh-LTB

β-Ac�n

β-Ac�n β-Ac�n β-Ac�n

LTβ

resistanceCetuximabsignature (n = 170 iden��es,

Table S1B)

Genes expression ≧2.5 folds

with FPKM>1 in FaDu-Snail cells (n = 477 genes,

Table S1C)

Snail-regulated CtxR candidates (n=13 genes, IL7R, LTB, ETS1, SERPINE1, ARL4C, MAP4K4, LIMA1, ICAM1,CSRP2, TBC1D16, TNFAIP3,

SEC24D, NFKBIE)

Top 10% leading edge candidates

for cetuximabresistance

(Suppl. Fig. S2A): LTB

Genes expression≧1.5 folds

in CtxR cells from GSE21483

(n = 1,548 iden��es)

Genes expression≧1.5 folds

in OECM1-CtxR cells(n = 1,955 iden��es,

Table S1A)

0 sh-scr sh-LTB

**

Rela

�ve

inhi

bi�o

n (%

)

FaDu-Snail

2040

6080

sh-scr sh-LTB

Snail

LTβ

FaDu-Snail

6543210

Non Rec

P = 0.046

Rel

ativ

e LTB

expr

essi

on

Figure 2.

LTb is the major target of Snail to mediate cetuximab resistance. A, Flowchart showing the mining process of the Snail-regulated target genes for mediatingcetuximab resistance. B, Western blot analysis of LTb in 2 cetuximab-resistant sublines versus control (OECM1-CtxR vs. OECM1-WT and FaDu-CtxR vs.FaDu-Ctrl).C,Histogram showing the result of relative expression of LTB in 12 cetuximab-treated patientswithHNSCC (4with and 8without recurrence). The result ispresented as the fold change of LTB expression in tumors with related to the corresponding normal counterparts. For calculating the result, the data fromboth tumor and normal tissues were normalized with the internal control GAPDH first and then the tumor-to-normal fold changes of LTBwere obtained. Non, non-recurrence (n¼ 8); Rec, recurrence (n¼ 4). Data represent mean� SD. P value is estimated by the Student t test.D, Left,Western blot analysis of LTb in OECM1-CtxR

(top), FaDu-CtxR (middle), and primary HNSCC cells harvested from a cetuximab-resistant patient (bottom) transfected with an shRNA against LTB (sh-LTB)or a scrambled control (scr). Right, Relative growth inhibition by cetuximab of the corresponding cell lines. Data represent mean � SD. n ¼ 3. �� , P < 0.01(Student t test). E, Scatter plot showing the correlation between the relative expression of LTB and SNAI1 in 58 patients with HNSCC from Taipei Veterans GeneralHospital (top) or 542 HNSCC samples from TCGA database. P value and correlation coefficient r are shown in each panel. F, Left, Western blot analysis of Snailand LTb in FaDu cells transfected with a Snail-expressing vector (FaDu-Snail) versus a control vector (FaDu-vec). Right, Western blot analysis of Snail and LTb inOECM1 cells receiving a shRNA against LTB (sh-LTB) or a scrambled control. G, Left, Western blot analysis of Snail and LTb in FaDu-Snail cells receiving ashRNA against LTB (sh-LTB) or a scrambled control (scr). Right, Relative growth inhibition by cetuximab of the corresponding cell lines. Data represent mean� SD.n ¼ 3. �� , P < 0.01 (Student t test).

Hsu et al.

Clin Cancer Res; 23(15) August 1, 2017 Clinical Cancer Research4392

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

0

20

40

60

80

0

1

2

3

4

5

6

7

C

A

D

F

Snail

LTβ

β-Ac�n

SnailVecFaDu

Snail2R

LUC

LUC

pGL4.2-LTB-WT

pGL4.2-LTB-MUT

Perc

enta

geof

inpu

t

CDH1 LTB_1 LTB_2

**

**

**SnailIP:B

CDH1 LTB-1 LTB-2

IgGIP:10

1086420

0

8642

0

1

2

3

Cell: FaDu

PGL4.1-LTB-MUT**

Vec

Snail+ +

+ +

PGL4.2-LTB-WT*

PGL4.2-LTB-MUT

Fold

chan

geof

rela

�ve

luci

fera

seac

�vity

0

1

2

3

0

1

2

3

Cell: 293T

PGL4.2-LTB-WT PGL4.2-LTB-MUT

**

Vec

Snail+ ––

––

––

– –––

–

– –

–– –

––

++ +

3

1

2

0

E

pCDHpCDH-SnailpHApHA-CBP

Cell: FaDu

**

*

Fold

chan

geof

rela

�ve

luci

fera

seac

�vity

++

++

++

IgG ac-SnailLys146

ac-SnailLys187

OECM1-WT

OECM1-CtxR

**

Perc

enta

geof

inpu

t

*

TSS-874 -773 -626 -135 -62

LTB-1 ChIP LTB-2 ChIP

-883E E EE

-841EE E

+394+268EE

+475–1008 –914 –759 –19–178

OECM1-CtxROECM1-WT

OECM1-WTOECM1-CtxR

Figure 3.Direct regulation of LTB transcription by acetylated Snail. A, Schematic representation of the proximal promoter region of LTB and reporter constructs used intransient transfection assays. E, E-box. The amplified region in ChIP assay in (B) and (F) is also indicated. pGL4.2-LTB-WT: wild-type LTB promoter construct;pGL4.2-LTB-MUT: E-boxes mutated LTB promoter construct. B, ChIP assay in OECM1-CtxR versus OECM1-WT. The percentages of enrichment were normalized toinput. Amplifying CDH1 promoter was a positive control for ChIP assay of Snail-overexpressed cells, and IgG was a control for immunoprecipitation. Datarepresent mean � SD. n ¼ 3. �� , P < 0.01 (Student t test). C, Luciferase reporter assay. Left two panels, Transient transfection of the wild-type/mutant reportertogether with Snail expression vector/control vector (Vec) and pCMV-b-gal into HEK-293T cells. Right two panels, Transfection of the wild-type/mutantreporter together with pCMV-b-gal into FaDu-Snail/FaDu-control stable transfectants. The luciferase activity values were normalized by the b-galactosidaseactivity. Data represent mean� SD. n¼ 3. � , P < 0.05; �� , P < 0.01 (Student t test). D,Western blot analysis of Snail and LTb of FaDu cells transfected with wild-typeSnail, unacetylatable Snail (Snail2R), or a control vector (Vec). E, Luciferase reporter assay. The FaDu cells were co-transfected with pGL4.2-LTB-WT, the indicatedexpression/control vectors, and pCMV-b-gal. The luciferase activity values were normalized by the b-galactosidase activity. Data represent mean � SD. n ¼ 3.� , P < 0.05; �� , P < 0.01 (Student t test). F, ChIP assay in OECM1-CtxR versus OECM1 parental cells. The cell lysates were immunoprecipitated by the antibodies thatspecifically recognize lysine 146- or lysine 187-acetylated Snail antibody, and LTB-2 primer in (A) was used to amplify the Snail-binding region. The percentages ofenrichment were normalized to input. IgG was a control for immunoprecipitation. Data represent mean � SD. n ¼ 3. � , P < 0.05; ��, P < 0.01 (Student t test).

Lymphotoxin-b Induces Cetuximab Resistance

www.aacrjournals.org Clin Cancer Res; 23(15) August 1, 2017 4393

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

C

D

A Cold ProbeOECM1-WT OECM1-CtxR

--- x30 x90- + - --

-- + + +

J

B

α-Tubulin

p65

p52

H3

C NFaDu

WT CtxR WT CtxR

C NOECM1

Ctrl CtxR Ctrl CtxR

H

p-IKKα

p-Akts473

sh-scr -LTB

LTβ

Akt

β -Ac�n

OECM1-CtxR

p-EGFRY1068

p-STAT3

STAT3

EGFR

β -Ac�n

OECM1-CtxR

LTβsh-scr -LTB

+ -Bortezomib +Cetuximab - + +

01020304050

E

Rela

�ve

inhi

bi�o

n (%

)

*

0

**

Cell: OECM-CtxR

+ -Parthenolide +Cetuximab - + +

20304050

10

****

Rela

�ve

inhi

bi�o

n (%

)

OECM-CtxR

Erb-iCetuximab

**50403020100

+ - + - + +

Rela

�ve

inhi

bi�o

n (%

)

Primary HNSCC

Erb-iCetuximab

**50403020100

+ - + - + +

NF- κ binding B

-3.0 -1.5 0 1.5 3.0

NFKB2N

Fκ B

EMT

IL1B

LTB

CCL5TNFACXCL10

CXCL1

CD44IL6

ICAM1

SNAI1VIMFN1

CDH1

OEC

M1-

CtxR

OEC

M1-

WT

I

G

p-IKKα

β-Ac�n

WT CtxR

IKK α

OECM1

EGFR

p-EGFRY1068

p-ERK1/2

ERK1/2

STAT3

p-STAT3

CtxRWTOECM1

β-Ac�n

LTB scrscrsh LTB

C N

p65

p52

TBP

Cell: OECM1-CtxR

α -Tubulin

LTBLTB scrscr

C N

Cell: FaDu-CtxR

0

10

20

30

40

Cell: FaDu-CtxR

BortezomibCetuximab

+−+ + + −

01020304050

ParthenolideCetuximab

+−+ + + −

Rel

ativ

ein

hibi

tion

(%)

****

*

**

Rel

ativ

ein

hibi

tion

(%)

F

1 3.3

1 4.2

Akt

p-AktS473

1 2.3

1 1.5

1 5.1

1 0.4

1 0.57

p-ERK 1/2

ERK 1/2

1 0.8

1 0.45

p-Akts473

p-ERK 1/2EGFR

p-EGFR

Akt

ERK 1/2

FaDu-CtxR

sh-scr -LTB

STAT3

p-STAT3

β -Ac�n

LTβ

sh-scr -LTBFaDu-CtxR

1 0.15

1 0.05

1 0.45

1 0.95

β-Ac�n

Clin Cancer Res; 23(15) August 1, 2017 Clinical Cancer Research4394

Hsu et al.

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

of the upregulated genes from the microarray of OECM1-CtxR

cells versus OECM1 parental cells (Supplementary Table S1A)with the genes identified from the RNA-seq for cetuximab resis-tance (GSE21483, ref. 26). We next examined the overlap of thecore signature of resistance (Supplementary Table S1B) and thesignature from the RNA-seq for FaDu-Snail/FaDu-vector (Sup-plementary Table S1C). A total of 13 upregulated genes werefound as candidates. Among these candidates, LTB, whichencodes LTb, was the only one in the top 10% of the "leading-edge core genes for cetuximab resistance", that is, the only genethat ranked in the top 10% of the cetuximab-resistant signaturegenes (Fig. 2A; Supplementary Fig. S2A). We therefore focusedon LTb in this study. LTb is a TNF family membrane protein andplays amajor role in lymphoid cells (27, 28), but its role in cancercells has been investigated to a limited extent. Increased LTbexpression was observed in both cetuximab-resistant sublines(Fig. 2B).Higher expressionof LTBwasdemonstrated in samplesfrom patients with HNSCC who developed recurrence aftercetuximab treatment (Fig. 2C). Knockdown of LTb increasedcetuximab sensitivity in both OECM1-CtxR and FaDu-CtxR cellsand also sensitized the primary HNSCC cells obtained from acetuximab-resistant patient (Fig. 2D). We next investigatedthe correlation between Snail and LTb in HNSCC. A positivecorrelation between the expression of SNAI1 and LTB was dem-onstrated both in 58 HNSCC samples from Taipei VeteransGeneral Hospital and in 541 HNSCC samples from TCGAdatabase (Fig. 2E). A positive correlation between Snail and LTbwasnoted inHNSCCcell lines (Supplementary Fig. S2B). EctopicSnail upregulated LTb, whereas knockdown of Snail suppressedLTb (Fig. 2F). Induction of SNAI1 expression by TGF-b inOECM1 cells upregulated LTB, and knockdown of Snail abro-gated this effect (Supplementary Fig. S2C). Knocking down LTbin FaDu-Snail cells attenuated Snail-induced cetuximab resis-tance (Fig. 2G).

We next investigated the relationship between LTb and Snail,especially paid attention to observe whether a reciprocal reg-ulation between Snail and LTb exists. Knockdown of LTb inOECM1 and SAS cells did not reduce Snail correspondingly;consistently, ectopic expression of LTb in FaDu did not upre-gulate Snail (Supplementary Fig. S2D), suggesting that LTb is

located downstream to Snail. Next, we investigated whetherLTb itself is able to induce EMT in HNSCC because EMT isassociated with cetuximab resistance. Knockdown of LTb inOECM1-CtxR cells reversed the mesenchymal phenotype (Sup-plementary Fig. S2E), and overexpression of LTb in FaDu cellspromoted EMT (Supplementary Fig. S2F and S2G). In summa-ry, the above findings suggest that LTb is the key molecule inmediating Snail-induced cetuximab resistance in HNSCC, andLTb itself is able to induce EMT.

Acetylated Snail induces LTb expressionWe investigated whether Snail induces LTb through tran-

scriptional activation. A chromatin immunoprecipitation(ChIP) assay demonstrated increased binding of Snail on theregion of the LTB proximal promoter which harbors the Snail-binding sites in OECM1-CtxR cells compared with parentalOECM1 cells (Fig. 3A and B). Mutation of the two E-boxes onthe LTB promoter abrogates Snail-induced LTB promoter acti-vation (Fig. 3C). Because we previously showed that CBP-mediated acetylation of Snail at lysine 146 and lysine 187switches its function from a transcriptional repressor to anactivator (29), we investigated whether acetylated Snail acti-vates LTB transcription. Transfection of an unacetylatable Snailmutant (Snail2R) induced a lesser degree of LTb expressionthan the wild-type Snail did (Fig. 3D). Co-transfection of CBPwith Snail augmented the Snail-induced LTB promoter activa-tion (Fig. 3E). Increased binding of acetylated Snail to the LTBpromoter was observed in OECM1-CtxR cells (Fig. 3F). Togeth-er, the above data indicate that acetylated Snail promotes LTBtranscription.

LTb activates NF-kB signal pathways and increases EGFRphosphorylation in cetuximab-resistant cells

In lymphoid cells, LTb forms a membrane-anchored hetero-trimer with lymphotoxin-a (LTa; refs. 30, 31), and LTa1b2heterotrimer binds to lymphotoxin-b receptor (LTbR) to acti-vate the NF-kB pathways (32, 33). We herein investigatedwhether LTb activates NF-kB to mediate cetuximab resistance.Analyzing the microarray data obtained from OECM1-CtxR

versus OECM1 parent cells revealed activation of NF-kB target

Figure 4.Activation of NF-kB pathway in cetuximab-resistant cells.A,Heatmap summarizing the results of the expression of NF-kB target genes and EMT-related genes fromthe cDNAmicroarray data of OECM1-CtxR versus OECM1-WT. B,Western blot analysis of the nuclear and cytoplasmic fractions of p65 and p52 in OECM1-CtxR versusOECM1-WT (left) and FaDu-CtxR versus FaDu-Ctrl. C, cytoplasm; N, nucleus. a-Tubulin was a loading control for cytoplasmic fraction, and H3 was a controlfor nuclear fraction. C, Electrophoretic mobility shift assay. The NF-kB activity of OECM1-WT and OECM1-CtxR cells was detected by incubating the nuclear extractswith the biotin-labeled DNA probe containing the NF-kB–binding sequence. Different folds of unlabeled probes were used for competition assay. The arrowindicates the shifted band of NF-kB. D, Western blot analysis of the nuclear and cytoplasmic fractions of p65 and p52 in OECM1-CtxR (left) or FaDu-CtxR (right)transfected with a shRNA against LTB or a control sequence. C, cytoplasm; N, nucleus. a-Tubulin was a loading control for cytoplasmic fraction, and TBPwas a control for nuclear fraction. E, Left, Relative growth inhibition by cetuximab of OECM1-CtxR cells treated with parthenolide, cetuximab, or both. Right, Relativegrowth inhibition by cetuximab of OECM1-CtxR cells treated with bortezomib, cetuximab, or both. Working concentration of parthenolide: 15 mmol/L,bortezomib: 20 nmol/L, cetuximab: 300 mg/mL. Data represent mean � SD. n ¼ 3. �� , P < 0.01 (Student t test). F, Top, Relative growth inhibition by cetuximab ofFaDu-CtxR cells treated with parthenolide, cetuximab, or both. Bottom, Relative growth inhibition by cetuximab of FaDu-CtxR cells treated with bortezomib,cetuximab, or both.Working concentration of parthenolide: 15mmol/L, bortezomib: 20 nmol/L, cetuximab: 300 mg/mL. Data represent mean� SD. n¼ 3. � , P < 0.05;�� , P < 0.01 (Student t test). G, Western blot analysis of phosphor-IKKa/total IKKa, phosphor-EGFR/total EGFR, phosphor-Akt/total Akt, phosphor-Erk/totalErk, and phosphor-STAT3/total STAT3 in OECM1-CtxR versus OECM1-WT. H, Western blot analysis of LTb, phosphor-IKKa, phosphor-EGFR Y1068/total EGFR,phosphor-Akt/total Akt, phosphor-ERK/total ERK, and phosphor-STAT3/total STAT3 in OECM1-CtxR receiving an shRNA against LTB or a control sequence. I,Western blot analysis of LTb, phosphor-EGFR Y1068/total EGFR, phosphor-Akt/total Akt, phosphor-ERK/total ERK, and phosphor-STAT3/total STAT3 inFaDu-CtxR receiving an shRNA against LTB or a control sequence. J, Left, Relative growth inhibition by cetuximab of OECM1-CtxR cells treated with 3 mmol/L Erbreceptor inhibitor JNJ28871063 hydrochloride (Erb-i), cetuximab (300 mg/mL), or both. Right, Relative growth inhibition by cetuximab of primary HNSCC cellstreated with JNJ28871063 hydrochloride (3 mmol/L), cetuximab (1000 mg/mL), or both. Data represent mean � SD. n ¼ 3. �� , P < 0.01 (Student t test).

www.aacrjournals.org Clin Cancer Res; 23(15) August 1, 2017 4395

Lymphotoxin-b Induces Cetuximab Resistance

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

genes as well as of an EMT signature in OECM1-CtxR cells(Fig. 4A). NF-kB activation was evident by an increased expres-sion of nuclear p65 and p52 proteins in both resistant sublines(Fig. 4B). An enhanced NF-kB activity in OECM1-CtxR wasconfirmed by electrophoretic mobility shift and reporter assays(Fig. 4C and Supplementary Fig. S3A). Knocking down LTBreduced nuclear p65 and p52 in 2 resistant sublines (Fig. 4D),and suppression of LTB downregulated the NF-kB reporteractivity (Supplementary Fig. S3B). Knockdown of LTb alsoattenuated the expression of both canonical and noncanonicalNF-kB target genes (Supplementary Fig. S3C). Inhibition of NF-kB activity by parthenolide or bortezomib sensitized OECM1-CtxR or FaDu-CtxR to cetuximab (Fig. 4E and F), and primaryHNSCC cells also confirmed this result (SupplementaryFig. S3D). An increased TNFa level was observed in the super-natant of OECM1-CtxR compared to that of the parentalOECM1 cells (Supplementary Fig. S3E), suggesting a crucialrole of TNFa in maintaining NF-kB pathway activation incetuximab-resistant cells. Because LTb has been reported to bea target induced by NF-kB (34), we examined whether sup-pression of NF-kB activity also downregulates LTb. The resultshowed that treatment of OECM1-CtxR with either partheno-lide or bortezomib downregulated mRNA level of LTB as well asthe NF-kB targets CCL2 and P100 (Supplementary Fig. S3F).The protein level of LTb was also suppressed by the NF-kBinhibitor in OECM1-CtxR (Supplementary Fig. S3G).

Next, we investigated whether LTb activates NF-kB through anLTa/LTbR-dependent mechanism (32, 33). To this end, we exam-ined the expression of LTa and LTbR in cetuximab-resistant cells.While the expression of LTb and LTbR was evident in OECM1-CtxR, these cells barely expressed LTa (Supplementary Fig. S4A–S4C). We therefore considered that LTb activates NF-kB throughan LTa/LTbR-independent pathway in cetuximab-resistantHNSCC. Because constitutive activation of EGFR is frequentlynoted in anti-EGFR–resistant cancer cells and Akt, a major down-stream kinase of EGFR, has been shown to induce NF-kB activa-tion through directly phosphorylating IKKa (35) or through anmTOR-mediated mechanism (36), we examined whether EGFR-mediated Akt activation is responsible for NF-kB activation incetuximab-resistant cells. An increased EGFR phosphorylation attyrosines 845, 1068, 1086, and 1148 was shown in both resistantsublines (Supplementary Fig. S4D), and higher levels of phos-phorylated Akt, ERK, IKKa, and STAT3 were also noted (Fig. 4G),which suggests that the EGFR/Akt/IKKa/NF-kB pathway is acti-vated in cetuximab-resistant cells. Interestingly, knocking downLTb in two resistant sublines inhibited the phosphorylation ofEGFR and its downstream signal molecules including Akt, IKKa,and STAT3 (Fig. 4H and I), suggesting a potential crosstalkbetween LTb and the EGFR/Akt/IKKa-NF-kB pathway. The EGFRtyrosine kinase inhibitor afatinib and the PI3K inhibitorLY294002 each partially reversed the cetuximab resistance–induced expression of NF-kB target genes (Supplementary Fig.S4E). Inhibition of EGFR activity sensitized OECM1-CtxR and theprimary HNSCC cells to cetuximab treatment (Fig. 4J). Together,the results suggest that LTb activatesNF-kB through the EGFR/Akt/IKKa pathway.

Afatinib and bortezomib exert a synergistic effect in cetuximab-resistant HNSCC cells

Because inhibition of EGFR or Akt activation only partiallysuppressed NF-kB target gene expression in cetuximab-resistant

cells (see Supplementary Fig. S4E), we examined whether com-bining the EGFR tyrosine kinase inhibitor afatinib and the NF-kBinhibitor parthenolide or bortezomib, which is considered anNF-kB inhibitor (37), would be more effective. The combinationof afatinib and bortezomib provided a stronger inhibitoryeffect for both resistant sublines (Fig. 5A). A synergistic effect byexamining the combination index was demonstrated betweenafatinib/bortezomib and afatinib/parthenolide (Fig. 5B). Wenext utilized the zebrafish model to confirm their synergy invivo. Each 2-day-old zebrafish embryo was injected with 100labeled OECM1-CtxR cells. The embryos were treated withdifferent combinations of drugs 1 day after injection, and thetumor cells were observed and recorded one day (baseline) orthree days after tumor cell injection. A significantly increasedinhibitory effect on OECM1-CtxR proliferation was shown inthe afatinib plus bortezomib group compared with afatinibalone or afatinib combined with other drugs (Fig. 5C). Thesedata indicate a synergistic effect of afatinib and bortezomib incetuximab-resistant HNSCC cells.

LTb interacts with methylated EGFR to enhance EGFRdimerization and activation

We next investigated the mechanism of LTb-induced EGFRphosphorylation. First, we examined whether LTb interactswith EGFR. In FaDu-CtxR, endogenous LTb interacted withEGFR (Fig. 6A). Ectopic LTb also interacted with EGFR (Sup-plementary Fig. S5A). A proximity ligation assay confirmedthe interaction between LTb and total/phosphorylated EGFRin another cetuximab-resistant subline OECM1-CtxR (Supple-mentary Fig. S5B). LTb preferentially interacted with the extra-cellular domain (ECD) of EGFR (Supplementary Fig. S5C).We next investigated the impact of the LTb–EGFR interactionon EGFR ligand binding and dimerization. Significantlyreduced EGF–EGFR ligand binding after cetuximab treatmentwas shown in OECM1 parental cells compared with OECM1-CtxR (Fig. 6B). A reduced dissociation constant (Kd) for EGFRligand binding was noted after transfection with LTb (Supple-mentary Fig. S5D). Because ligand binding–induced EGFRdimerization is responsible for EGFR activation (38), wefurther examined whether LTb affects EGFR dimerization.Increased EGFR dimerization was shown in OECM1-CtxR

(Supplementary Fig. S5E). Transfection of LTb promoteddimerization of wild-type EGFR. Mutation of valine 924of EGFR, a critical site mediating the formation of the EGFRdimer (39), reduced LTb-induced EGFR dimerization (Fig.6C). We previously found that protein arginine methyltrans-ferase 1 (PRMT1)–mediated methylation of EGFR R198/R200 increases EGFR dimerization and ligand binding, result-ing in cetuximab resistance (13). As the LTb-induced pheno-type is very similar to that resulting from EGFR methylation,we investigated the crosstalk between these two mechanisms.Increased R198/R200 methylated EGFR and PRMT1 wasobserved in OECM1-CtxR (Supplementary Fig. S5F). EctopicSnail increased methylated EGFR and PRMT1 in FaDucells (Supplementary Fig. S5G), and knocking down endoge-nous Snail in OECM1 cells downregulated PRMT1 significant-ly (Fig. 6D). Importantly, LTb preferentially interacted withmethylated EGFR (Fig. 6E). These results indicate that Snailincreases EGFR methylation by upregulating PRMT1, whichfacilitates the EGFR–LTb interaction to induce cetuximabresistance.

Clin Cancer Res; 23(15) August 1, 2017 Clinical Cancer Research4396

Hsu et al.

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

C

A

76% 68%

33%50%50%50%

12%16%

33%20% 25% 28%

12% 16%33% 30% 25% 22%

0%10%20%30%40%50%60%70%80%90%

100%Cell number increase Cell number stay the same Cell number decreased

1dpi

3dpi

Ki67

a

a’

b

b’

c

c’

d

d’

e

e’

f

f’

g jih

B

OECM1-CtxR

PBS

OECM1-CtxR

Afa�nib

OECM1-CtxR

+Afa�nibBortezomib

OECM1-CtxR

+Afa�nibLY294002

OECM1-CtxR

+Afa�nibRapamycin

OECM1-CtxR

+Afa�nibS31-201

OECM-CtxR

2

1

0

–1

–20 0.5 1

Fa

Log(

CI)

+Afa�nib Bortezomib 2

1

0

–1

–20 0.5 1

Parthenolide+Afatinib

OECM-CtxR FaDu-CtxR

CetuximabAfa�nib

Bortezomib

+ - --- + - +

-- + +

+ - --- + - +

-- + +

01020

304050

**

**Re

la�v

ein

hibi

�on

(%)

0102030

4050

**

**

Fa

Figure 5.

A combinatory effect of afatinib and bortezomib in cetuximab-resistant HNSCC cells.A,Relative growth inhibition by cetuximabofOECM1-CtxR cells (left) and FaDu-CtxR cells (right) treated with cetuximab (300 mg/mL), afatinib (15 mmol/L), bortezomib (20 nmol/L), or afatinib þ bortezomib. Data represent mean � SD.n ¼ 3. �� , P < 0.01 (Student t test). B, Combination index (CI) plot for showing the synergy of afatinib and bortezomib (left) or afatinib and parthenolide (right) inOECM1-CtxR. Fa, fraction affected.C,Quantification ofOECM1-CtxR engrafted andproliferated in the zebrafish xenotransplantationmodelwithout drug (PBS) orwithdifferent combination of drugs. Proliferation inhibition was determined by comparing the intensity of CFSE-labeled cells at 3 dpi (after drug treatment)versus 1 dpi (before drug treatment). The percentages of embryos carrying increase cell number are shown as red, same cell number is denoted as gray, anddecreased cell number is indicated as green. Representative fluorescent images of zebrafish embryos transplanted with OECM1-CtxR without drug at 1 dpi (a) and 3dpi (a'); with afatinib at 1 dpi (b) and 3 dpi (b0); with afatinibþbortezomib at 1 dpi (c) and 3 dpi (c0); with afatinibþ LY294002 at 1 dpi (d) and 3dpi (d0); with afatinibþrapamycin at 1 dpi (e) and 3 dpi (e0) and afatinib þ S3I-201 at 1 dpi (f) and 3 dpi (f0). Each embryo was xenotransplanted with approximately 100 cells.Groups of 20 embryos were used for each drug treatment. Scale bars are 1.0 mm. The sacrificing embryos at 3 dpi were performing Ki67 immunostaining to verifythe proliferation. Representative bright-field (G) and CFSE fluorescent image (H), Ki67-positive image (I) as well as overlapping of CFSE and Ki67-positiveimage (J) of zebrafish embryos transplanted with OECM1-CtxR treated with afatinib. Ki67 immunostaining of the 3 dpi embryos indicated the CFSE stainingco-localized with the Ki67 staining indicated the CFSE positive cells are proliferated cells.

www.aacrjournals.org Clin Cancer Res; 23(15) August 1, 2017 4397

Lymphotoxin-b Induces Cetuximab Resistance

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

A C

E F

IP: Input IgG LTβ

EGFR

LTβ

Cell: FaDu-CtxR

IB

D2

D1

D3

β1 β2/3 β4 β5 β6 β7 β8 β9 β10

EGFR

EGFRECD

250

200

150

100

50

0CetuximabLTβ pep�de

+ + + +Mock β5 β8 β9

G H

- + - + - +BS3

WT WT V924RLTB

EGFRdimer

EGFRmonomer

EGFR

- - + + + +

OECM1

sh-scr sh-Snail

PRMT1

Snail

β-Ac�n

D

B

75.9%

42%

EGF-FITC

OECM1-CtxR

OECM1-WT

OECM1-CtxR

+ Cetuximab

OECM1-WT+Cetuximab

10.5%

43.7%

EGFR WT R198K/R200K

IP: FLAGIB: HAIP: FLAGIB: FLAG

FLAG

HAInpu

t

Cell: 293T+pHA-EGFR+pFLAG-LTB+pCMV5-PRMT1

SSC

No.

ofco

loni

es

** **

0

10

20

30

40

50

+ + + +Mock β5 β8 β9

Cell: FaDu-CtxR

** *

***No.

of c

olon

ies

Cell: OECM1-CtxR

I

Clin Cancer Res; 23(15) August 1, 2017 Clinical Cancer Research4398

Hsu et al.

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

Finally, we investigatedwhether interfering with the EGFR–LTbinteraction is a potential strategy for reversing cetuximab resis-tance. Structurally, LTb is composed of 10 b-strands (b1–b10).Computational simulation of the molecular docking of EGFRand LTb showed that the b-strands of LTb are close to the D1region of EGFR–ECD, suggesting that the b-strands of LTb mayserve as an interface for the EGFR–LTb interaction (Fig. 6F). Wenext synthesized nine peptides with the sequences of LTbb-strands for investigating the effect of intercepting LTb–EGFRinteraction in cetuximab-resistant cells (Supplementary Fig.S5H). The in vitro binding assay showed that the peptidesb5, b8, and b9 interacted with EGFR–ECD (Fig. 6G). Bothb5 and b9 suppressed EGFR Y1068 phosphorylation in tworesistant sublines and b9 had a more prominent effect (Sup-plementary Fig. S5I). b9 inhibited EGFR dimerization inOECM1-CtxR (Supplementary Fig. S5J). Furthermore, b9 alsoreduced EGF–EGFR binding (Supplementary Fig. S5K), and b5and b9 sensitized both OECM1-CtxR and FaDu-CtxR to cetux-imab treatment (Fig. 6H).

We propose a model to summarize our findings in Fig. 6I.Cetuximab-resistant HNSCC cells undergo a Snail-induced EMT.Snail increases the levels of both LTb and methylated EGFR. Theinteraction of LTb with methylated EGFR promotes EGFR dimer-ization, which activates the EGFR and NF-kB signaling pathwaysto induce cetuximab resistance. Interfering with the resistancepathway by using afatinib, bortezomib/NF-kB inhibitors, or theLTb 9th b-strand peptide reverses the resistance.

DiscussionA major finding of this study is that LTb interacts with

methylated EGFR to promote cetuximab resistance. We recentlyfound that methylation of EGFR at R198 and R200 by PRMT1enhances EGFR-ligand binding and EGFR dimerization, result-ing in cetuximab resistance (13). However, the inducing signalfor EGFRmethylation during the evolution of resistant clones iselusive. Here, we demonstrate that Snail acts as the key player incetuximab resistance by inducing both LTb and PRMT1, whichleads to the LTb-methylated EGFR interaction and downstreamsignal activation. The principal role of EMT in treatment resis-tance is therefore emphasized.

Here, we showed that acetylated Snail activates LTB tran-scription. Because we previously demonstrated that the majortargets induced by acetylated Snail are inflammatory-relatedcytokines including TNFa (29), and an elevated TNFa levelwas noted in cetuximab-resistant cells in this study, we suggestthat an inflammation-triggered EMT, such as TNFa- or IL6-induced EMT (40, 41), will be crucial for inducing acquiredresistance to cetuximab. Since a certain proportion of HNSCChas been considered as the "inflamed" tumors (42), ourfindings highlight the importance of tumor inflammation ineliciting cetuximab resistance in HNSCC. Interestingly, we alsofound that treatment of NF-kB inhibitor reduces the expres-sion of LTB in cetuximab-resistant cells, which is consistentwith the previous finding that LTB is a target gene of NF-kB(34). In cetuximab-resistant HNSCC, NF-kB is not only adownstream signal of constitutive EGFR activation, it alsoinduces LTb. However, for explaining the synergistic effectbetween the EGFR inhibitor and NF-kB inhibitor in cetuxi-mab-resistant HNSCC, the above finding cannot rule out thepossibility of Snail-mediated NF-kB signaling. Previous reportalso supports that Snail activates NF-kB pathway to conferchemoresistance (43). We therefore suggest that in HNSCCwith acquired resistance to cetuximab, Snail induces theexpression of LTb, which promotes constitutive activation ofEGFR pathways. In addition, Snail itself may activate NF-kBfor inducing resistance (Fig. 6I). A recent report shows themodest efficacy of using afatinib as the second-line treatmentin recurrent/metastatic HNSCC (44). Combining NF-kB inhi-bitors with afatinib to improve the effectiveness of EGFRtyrosine kinase inhibitors in treatment-resistant HNSCC maydeserve further clinical investigation.

In conclusion, our study reveals a driver gene mutation andcanonical RTK pathway–independent mechanism for acquiredresistance to cetuximab in HNSCC. This knowledge may beapplied to patients who develop multiple resistances withoutknown driver mutations. Combinatory therapy with an EGFRtyrosine kinase inhibitor and an NF-kB inhibitor will be effec-tive in LTb-mediated cetuximab resistance, and the possibilityof interfering with the interaction between LTb and EGFR byblocking peptides warrants further investigation for clinicalapplication.

Figure 6.LTb interactswithmethylatedEGFR topromoteEGFbinding, dimerization, and activation.A, Immunoprecipitation andWestern blot analysis showing the interactionbetween EGFR and LTb in OECM1-CtxR cells. IgG is a control for immunoprecipitation. B, Representative result of flow cytometry to show the EGF bindingof OECM1-CtxR versus OECM1-WT in the absence/presence of cetuximab (n ¼ 3). The percentage of EGF binding is shown in each panel. C, EGFR dimerizationassay in HEK-293T cells transfected with LTb, wild-type, or V924R mutant EGFR with/without the cross-linker BS3. The arrows indicate the bands of EGFRmonomer and dimer. D, Western blot analysis of Snail and PRMT1 in OECM1 cells receiving a shRNA against Snail or a scrambled sequence (scr). E,Immunoprecipitation and Western blot analysis showing the interaction between LTb and wild-type or R198/R200-mutated EGFR. The HEK-293T cells wereco-transfected with pFLAG-LTB, pHA-EGFR/pHA-EGFR-R198K/R200K, and pCMV5-PRMT1. F, Schema for presenting the molecular docking of EGFR-ECD(blue) and LTb (white). The EGFR-LTb interface is zoomed at right. The b-sheets of LTb are indicatedwith different colors.G,Peptide-binding assay. The purified full-length EGFR and EGFR-ECD was incubated with peptides containing sequences of different b-sheets of LTb. H, Clonogenicity assay. Left, OECM1-CtxR cells weretreated with peptides of different LTb b-sheets in the presence of cetuximab and then the cells were plated onto dishes for 10 days. Right, FaDu-CtxR

cells were treated with peptides of different LTb b-sheets in the presence of cetuximab and then the cells were plated onto dishes for 10 days. The colonies werestained with crystal violet and counted if the size was great than 0.5 mm. I, Schema showing the mechanism of acquired resistance to cetuximab in headand neck cancer cells. When the cancer cells undergo Snail-induced EMT, the methylated EGFR interacts with LTb, resulting in EGFR activation. Activated EGFRpromotes NF-kB activation through IKKa to engender cetuximab resistance. Snail may also directly activate NF-kB (dotted arrow). Inhibition of EGFRphosphorylation by afatinib, NF-kBactivation by bortezomib or other NF-kB inhibitors, and interception of the interaction between LTb andmethylated EGFRby thepeptide of ninth b-strand of LTb reverse the cetuximab resistance.

www.aacrjournals.org Clin Cancer Res; 23(15) August 1, 2017 4399

Lymphotoxin-b Induces Cetuximab Resistance

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: D.S.-S. Hsu, W.-L. Hwang, M.-C. Hung, M.-H. YangDevelopment of methodology: D.S.-S. Hsu, W.-L. Hwang, C.-H. Chu,P.-B. Chen, W.-H. HsuAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): W.-L. Hwang, Y.-H. Ho, H.-K. Lin, C.-Y. Lin,H.-Y. Lan, S.-K. TaiAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D.S.-S. Hsu, W.-L. Hwang, C.-H. Yuh, Y.-H. Ho,P.-B. Chen, H.-S. Lin, H.-K. Lin, S.-P. Wu, C.-Y. Lin, W.-H. Hsu, H.-Y. Lan,H.-J. Wang, S.-K. TaiWriting, review, and/or revision of themanuscript:D.S.-S. Hsu, W.-L. Hwang,M.-C. Hung, M.-H. YangAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): D.S.-S. Hsu, W.-L. Hwang, C.-H. Yuh,H.-J. Wang, S.-K. Tai, M.-C. HungStudy supervision: W.-L. Hwang, M.-C. Hung, M.-H. Yang

AcknowledgmentsWe thank BGI and MBGEN Biosciences for their generous assistance with

BGI Oseq-T service and bioinformatics analysis. We thank Dr. Cheng-ChiChang (National Taiwan University of Taiwan) for providing the SAS andCAL-27 cell lines and Dr. Kuo-Wei Chang (National Yang-Ming University of

Taiwan) for providing the OECM1 cell line. We thank Dr. Nien-Jung Chen(National Yang-Ming University of Taiwan) for providing the NF-kB reporterpGL4.32.

Grant SupportThis work was supported by Ministry of Science and Technology (105-2918-

010-002, 104-2321-B-010-005, and 103-2314-B-010-035 to M.-H. Yang; 105-2320-B-010-004 to D.S.-S. Hsu; 105-2320-B-038-009-MY2 to W.-L. Hwang),National Health Research Institutes (NHRI-EX105-10331BI to M.-H. Yang),Taipei Veterans General Hospital (V105C-069 toM.-H. Yang), Veterans GeneralHospitals-University System of Taiwan Joint Research Program (VGHUST105-G-4-1-2 to M.-H. Yang), a grant from Ministry of Education, Aim for the TopUniversity Plan (105AC-T201 to M.-H. Yang) a grant from Ministry of HealthandWelfare, Center of Excellence forCancer Research (MOHW105-TDU-B-211-134003 to M.-H. Yang), Cancer Prevention and Research Institute of Texas(RP150245 to M.C.H.), and Taipei Medical University (TMU104-AE1-B11 toW.-L. Hwang).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received August 3, 2016; revised August 31, 2016; accepted February 6, 2017;published OnlineFirst February 14, 2017.

References1. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer

statistics 2012. CA Cancer J Clin 2015;65:87–108.2. Grandis JR, Tweardy DJ. Elevated levels of transforming growth factor

and epidermal growth factor receptor messenger RNA are early mar-kers of carcinogenesis in head and neck cancer. Cancer Res 1993;53:3579–84.

3. Chung CH, Ely K, McGavran L, Varella-Garcia M, Parker J, Parker N, et al.Increased epidermal growth factor receptor gene copy number is associatedwith poor prognosis in head and neck squamous cell carcinomas. J ClinOncol 2006;24:4170–6.

4. Bonner JA, Harari PM, Giralt J, Azarnia N, Shin DM, Cohen RB, et al.Radiotherapy plus cetuximab for squamous-cell carcinomaof the head andneck. N Engl J Med 2006;354:567–78.

5. Vormorken JB, Mesia R, Rivera F, Remenar E, Kawecki A, Rottey S, et al.Platinum-based chemotherapy plus cetuximab in head and neck cancer. NEngl J Med 2008;359:1116–27.

6. Bonner JA, Harari PM, Giralt J, Cohen RB, Jones CU, Sur RK, et al.Radiotherapy plus cetuximab for locoregionally advanced head andneck cancer: 5-year survival data from a phase 3 randomised trial, andrelation between cetuximab-induced rash and survival. Lancet Oncol2010;11:21–8.

7. Rabinowits G, Haddad RI. Overcoming resistance to EGFR inhibitor inhead and neck cancer: a review of the literature. Oral Oncol 2012;48:1085–9.

8. Montagut C, Dalmases A, Bellosillo B, Crespo M, Pairet S, Iglesias M, et al.Identification of a mutation in the extracellular domain of the EpidermalGrowth Factor Receptor conferring cetuximab resistance in colorectalcancer. Nat Med 2012;18:221–3.

9. DeRoockW, JonkerDJ,DiNicolantonio F, Sartore-Bianchi A, TuD, Siena S,et al. Association of KRAS p.G13Dmutation with outcome in patients withchemotherapy-refractory metastatic colorectal cancer treated with cetux-imab. JAMA 2010;304:1812–20.

10. Yonesaka K, Zejnullahu K, Okamoto I, Satoh T, Cappuzzo F, Souglakos J,et al. Activation of ERBB2 signaling causes resistance to the EGFR-directedtherapeutic antibody cetuximab. Sci Transl Med 2011;3:99ra86.

11. Lu Y, Li X, Liang K, Luwor R, Siddik ZH, Mills GB, et al. Epidermal growthfactor receptor (EGFR) ubiquitination as a mechanism of acquired resis-tance escaping treatment by the anti-EGFR monoclonal antibody cetux-imab. Cancer Res 2007;67:8240–7.

12. Bardelli A, Corso S, Bertotti A, Hobor S, Valtorta E, Siravegna G, et al.Amplification of the MET receptor drives resistance to anti-EGFR therapiesin colorectal cancer. Cancer Discov 2013;3:658–73.

13. Liao HW, Hsu JM, Xia W, Wang HL, Wang YN, Chang WC, et al. PRMT1-mediated methylation of the EGF receptor regulates signaling and cetux-imab response. J Clin Invest 2015;125:4529–43.

14. Misales S, Yaeger R, Hobor S, Scala E, Janakiraman M, Liska D, et al.Emergence of KRAS mutations and acquired resistance to anti-EGFRtherapy in colorectal cancer. Nature 2012;486:532–6.

15. Langer CJ. Exploring biomarkers in head and neck cancer. Cancer 2012;118:3882–92.

16. Vermorken JB, Herbst RS, Leon X, Amellal N, Baselga J. Overview of theefficacy of cetuximab in recurrent and/or metastatic squamous cell carci-noma of the head and neck in patients who previously failed platinum-based therapies. Cancer 2008;112:2710–9.

17. Thiery JP, Acloque H, Huang RY, Nieto MA. Epithelial-mesenchymaltransitions in development and disease. Cell 2009;139:871–90.

18. Mani SA, Guo W, Liao MJ, Eaton EN, Ayyanan A, Zhou AY, et al. Theepithelial-mesenchymal transition generates cells with properties of stemcells. Cell 2008;133:704–15.

19. Fischer KR, Durrans A, Lee S, Sheng J, Li F, Wong ST, et al. Epithelial-to-mesenchymal transition is not required for lungmetastasis but contributesto chemoresistance. Nature 2015;527:472–6.

20. Zheng X, Carstens JL, Kim J, Scheible M, Kaye J, Sugimoto H, et al.Epithelial-to-mesenchymal transition is dispensable for metastasisbut induces chemoresistance in pancreatic cancer. Nature 2015;527:525–30.

21. Lamouille S, Xu J, Derynck R. Molecular mechanisms of epithelial-mes-enchymal transition. Nat Rev Mol Cell Biol 2014;15:178–96.

22. Skvortsova I, Skvortsov S, RajuU, Stasyk T, RiestererO, Schottdorf EM, et al.Epithelial-to-mesenchymal transition and cmyc expression are the deter-minants of cetuximab-induced enhancement of squamous cell carcinomaradioresponse. Radiother Oncol 2010;96:108–15.

23. WongYK,ChangKW,ChenCF,ChangCS.Characterization of twonew celllines derived fromoral cavity human squamous cell carcinomas–OC1 andOC2. J Oral Maxillofac Surg 1990;48:385–390.

24. De Roock W, Claes B, Bernasconi D, De Schutter J, Biesmans B, FountzilasG, et al. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on theefficacy of cetuximab plus chemotherapy in chemotherapy-refractory

Clin Cancer Res; 23(15) August 1, 2017 Clinical Cancer Research4400

Hsu et al.

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

metastatic colorectal cancer: a retrospective consortium analysis. LancetOncol 2010;11:753–62.

25. Bertotti A, Migliardi G, Galimi F, Sassi F, Torti D, Isella C, et al. Amolecularly annotated platform of patient-derived xenografts ("xenopa-tients") identifies HER2 as an effective therapeutic target in cetuximab-resistant colorectal cancer. Cancer Discov 2011;1:508–23.

26. Hatakeyama H, Cheng H, Wirth P, Counsell A, Marcrom SR, Wood CB,et al. Regulation of heparin-binding EGF-like growth factor by miR-212and acquired cetuximab-resistance in head and neck squamous cell car-cinoma. PLoS One 2010;5:e12702.

27. Browning JL, Ngam-ek A, Lawton P, DeMarinis J, Tizard R, Chow EP, et al.Lymphotoxin beta, a novel member of the TNF family that forms aheteromeric complex with lymphotoxin on the cell surface. Cell 1993;72:847–56.

28. Ware CF. Network communications: lymphotoxins, LIGHT, and TNF.Annu Rev Immunol 2005;23:787–819.

29. Hsu DS, Wang HJ, Tai SK, Chou CH, Hsieh CH, Chiu PH, et al. Acetylationof Snail modulates the cytokinome of cancer cells to enhance the recruit-ment of macrophages. Cancer Cell 2014;26:534–48.

30. Williams-Abbott L, Walter BN, Cheung TC, Goh CR, Porter AG, Ware CF.The lymphotoxin-alpha (LTalpha) subunit is essential for the assembly,but not for the receptor specificity, of the membrane-anchored LTalpha1-beta2 heterotrimeric ligand. J Biol Chem 1997;272:19451–6.

31. Browning JL, Sizing ID, Lawton P, Bourdon PR, Rennert PD, Majeau GR,et al. Characterizationof lymphotoxin-alpha beta complexes on the surfaceof mouse lymphocytes. J Immunol 1997;159:3288–98.

32. Vallabhapurapu S, Karin M. Regulation and function of NF-kappaBtranscription factors in the immune system. Annu Rev Immunol 2009;27:693–733.

33. Dejardin E, Droin NM, Delhase M, Haas E, Cao Y, Makris C, et al. Thelymphotoxin-beta receptor induces different patterns of gene expressionvia two NF-kappaB pathways. Immunity 2002;17:525–35.

34. Voon DC, Subrata LS, Karimi M, Ulgiati D, Abraham LJ. TNF and phorbolesters induce lymphotoxin-beta expression through distinct pathways

involving Ets and NF-kappa B family members. J Immunol 2004;172:4332–41.

35. Ozes ON, Mayo LD, Gustin JA, Pfeffer SR, Pfeffer LM, Donner DB. NF-kappaB activation by tumour necrosis factor requires the Akt serine-threonine kinase. Nature 1999;401:82–5.

36. Dan HC, Cooper MJ, Cogswell PC, Duncan JA, Ting JP, Baldwin AS. Akt-dependent regulation of NF-kB is controlled by mTOR and Raptor inassociation with IKK. Genes Dev 2008;22:1490–500.

37. Chen D, Frezza M, Schmitt S, Kanwar J, Dou QP. Bortezomib as the firstproteasome inhibitor anticancer drugs: current status and future perspec-tive. Curr Cancer Drug Targets 2011;11:239–53.

38. Schlessinger J. Ligand-induced, receptor-mediated dimerization and acti-vation of EGF receptor. Cell 2002;110:669–72.

39. Red Brewer M, Yun CH, Lai D, Lemmon MA, Eck MJ, Pao W., et al.Mechanism for activation of mutated epidermal growth factor receptorsin lung cancer. Proc Natl Acad Sci U S A 2013;110:E3595–604.

40. Li CW, Xia W, Huo L, Lim SO, Wu Y, Hsu JL, et al. Epithelial-mesenchymaltransition induced by TNF-a requires NF-kB-mediated transcriptionalupregulation of Twist1. Cancer Res 2012;72:1290–300.

41. Sullivan NJ, Sasser AK, Axel AE, Vesuna F, Raman V, Ramirez N, et al.Interleukin-6 induces an epithelial-mesenchymal transition phenotype inhuman breast cancer cells. Oncogene 2009;28:2940–7.

42. BonomiM, Patsias A, PosnerM, Sikora A. The role of inflammation in headand neck cancer. Adv Exp Med Biol 2014;816:107–27.

43. Lim S, Becker A, Zimmer A, Lu J, Buettner R, Kirfel J. SNAI1-mediatedepithelial-mesenchymal transition confers chemoresistance and cellularplasticity by regulating genes involved in cell death and stem cell main-tenance. PLoS One 2013;8:e66558.

44. Machiels JP, Haddad RI, Fayette J, Licitra LF, Tahara M, Vermorken JB,et al. Afatinib versus methotrexate as second-line treatment in patientswith recurrent or metastatic squamous-cell carcinoma of the head andneck progressing on or after platinum-based therapy (LUX-Head & Neck1): an open-label, randomised phase 3 trial. Lancet Oncol 2015;16:583–94.

www.aacrjournals.org Clin Cancer Res; 23(15) August 1, 2017 4401

Lymphotoxin-b Induces Cetuximab Resistance

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955

2017;23:4388-4401. Published OnlineFirst February 14, 2017.Clin Cancer Res   Dennis Shin-Shian Hsu, Wei-Lun Hwang, Chiou-Hwa Yuh, et al.   Resistance to Cetuximab in Head and Neck Cancer

Interacts with Methylated EGFR to Mediate AcquiredβLymphotoxin-

  Updated version

  10.1158/1078-0432.CCR-16-1955doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://clincancerres.aacrjournals.org/content/suppl/2017/02/14/1078-0432.CCR-16-1955.DC1

Access the most recent supplemental material at:

   

   

  Cited articles

  http://clincancerres.aacrjournals.org/content/23/15/4388.full#ref-list-1

This article cites 44 articles, 12 of which you can access for free at:

  Citing articles

  http://clincancerres.aacrjournals.org/content/23/15/4388.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.org

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/23/15/4388To request permission to re-use all or part of this article, use this link

on May 25, 2020. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 14, 2017; DOI: 10.1158/1078-0432.CCR-16-1955